Synthesized builds the stand-in data that lets software rehearse - realistic enough to catch bugs, synthetic enough to expose no one.
Every piece of enterprise software has to be tested before it ships. To test it properly, engineers need data that looks and behaves like the real thing - customer records, transactions, account histories. The problem is obvious: real data is sensitive, regulated, and slow to get hold of.
Synthesized removes that bottleneck. Its platform takes data from production systems and produces a governed, production-realistic stand-in - generated, masked, and subset so that it keeps the statistical shape of the original while stripping out the identifiable information. Teams can then build, test, and migrate against data that is safe to touch. The company's own framing is direct: "Validate software, migrations, and AI agents before they touch production."
Founded in 2020 by Nicolai Baldin, who holds a PhD in statistics and machine learning from the University of Cambridge, Synthesized grew out of a gap he noticed between what data science could do in a lab and what large organizations could actually use. The result is a test-data automation platform now used by regulated enterprises across banking, insurance, healthcare, energy, and telecom.
Industry research puts hard numbers on a quiet, expensive chore. Synthesized built a company around closing these gaps.
Figures cited by Synthesized and customer case studies. Treat as company-reported, not audited.
"We are making sure we really identify those things which are going to break your app - at the data level, on the environment level - and help you expose those breakage points."- Nicolai Baldin, CEO & Founder
Synthesized ships as software that plugs into existing data pipelines and CI/CD workflows, deployed on customer premises or on AWS, Azure, and Google Cloud - with no code transfer required.
Cloud-native test data management for automated provisioning, generation, masking, and subsetting - with CI/CD integration across SAP HANA, PostgreSQL, SQL Server, Oracle, MySQL, DB2, and Salesforce.
Machine-learning-focused synthetic data generation for model development, analytics, and privacy-preserving data science - the research heart of the platform.
An open-source tool for inspecting datasets and models for hidden bias, so fairness problems surface before they reach production.
The old choice in QA was a bad one: use imitation data that is fast but thin, or copy real data that is comprehensive but slow and risky. Synthesized' pitch is that you no longer have to pick. Its generated data aims for 95%+ statistical accuracy while completely randomizing identifiable information - fast enough to produce a million-transaction dataset in roughly ten minutes, against an industry norm once measured in months.
It also sits in a crowded field. Tonic.ai, Syntho, Gretel, and Mostly AI chase similar ground, and legacy masking tools from IBM and Perforce (Delphix) have long served the enterprise. Synthesized' wedge is the enterprise plumbing - SAP-native automation across SAP and non-SAP systems, deep CI/CD integration, and on-premise deployment that never moves a customer's code. It's positioned less as a data-science toy and more as release-pipeline infrastructure.
"The issue of how best to manage and use this invaluable resource is no longer confined to the IT department."- Nicolai Baldin, CEO & Founder
Synthesized sells B2B, on subscription, to large regulated organizations and the engineering, QA, and data teams inside them. Named customers and users include Deutsche Bank and UBS, with industries spanning banking and financial services, healthcare and life sciences, insurance, energy and utilities, and telecom and media. Distribution runs through cloud marketplaces including Google Cloud and Microsoft Azure.
The customer relationships run unusually deep. Deutsche Bank is both a paying user - it halved its test-data discovery time - and an equity investor in the 2025 Series A. UBS occupies the same dual seat. For an infrastructure company, having your buyers on the cap table is a strong signal that the product is embedded, not just trialed.
Baldin's Cambridge PhD work surfaces the gap between data science and enterprise practice.
Synthesized closes a £2.2M seed round two weeks before the UK lockdown, then rebuilds for remote deployment.
Masking, subsetting, and CI/CD integration harden the test-data management platform for regulated enterprises.
AI-powered data solutions land on Google Cloud Marketplace; SAP-native test-data automation is introduced.
Redalpine leads a $20M round with Deutsche Bank and UBS participating, funding expansion across North America and Europe.
It automates the creation of production-realistic test data - generating, masking, and subsetting data from sensitive systems so teams can test software, migrations, and AI agents safely and in line with regulations like GDPR.
It was founded in 2020 by Nicolai Baldin, who holds a PhD in statistics and machine learning from the University of Cambridge and serves as CEO.
Publicly reported rounds include a £2.2M seed in 2020 and a $20M Series A in September 2025 led by Redalpine, with total funding reported around $26M+.
Large regulated enterprises including Deutsche Bank and UBS, across banking, insurance, healthcare, energy, telecom, and the public sector.
It combines synthetic data generation with masking and subsetting in one CI/CD-integrated platform, aiming to preserve 95%+ statistical accuracy while removing identifiable information - rather than only obscuring existing records.